We propose to screen 250 potential participants (or 1,250 across 5 sites). We will use two sources: medical referrals (n=125) and elementary school students (n=125). We expect to perform a comprehensive assessment of 200 individuals, which will include a structured interview, and to confirm a DSM-IV diagnosis of ADHD in 170 individuals who will enter the study as subjects. Subgroups will be formed based on presence of comorbid childhood disorders. These 170 subjects will be offered a double-blind, dose response assessment of 2 stimulant medications (methylphenidate and d- amphetamine), and we anticipate that 10% will either refuse or manifest an adverse response. This will leave approximately 150 stimulant-responsive ADHD to be randomly assigned to 1 of 5 treatment groups: Community Treatment (COM, n+30), Structured Medication (MED, n=30), MED plus School- Based intervention (MED+SCH, n=30), MED+SCH plus Parent Training (MED+SCH+PAR, n=30), and MED+SCH+PAR plus tailored multimodality treatment selected from the Satterfield menu (MED+SCH+PAR+SAT, n=30). If this were duplicated across 5 sites, this would yield a group size of 150 subjects in each of the 5 treatment groups. A 2 year treatment protocol will be implemented. We expect a 10% yearly attrition rate, so at the follow-up 3 years after treatment we expect to retain about 22 subjects (or 110 subjects across 5 sites) in each group. Two types of outcome measures will be taken to evaluate effects of treatments. First, we will use the IOWA and SNAP parent and teacher rating scales to obtain quarterly assessments of I/O and A/D behavior and ADHD symptomology to document short-term effects and progress in treatment. Second, we will perform a full assessment yearly to measure initial, maintenance, and long-term effects in 5 areas: home behavior, school behavior, school performance, peer status, and family functioning. Our sample size will give us adequate statistical power to evaluate the impact of multimodality treatment (MED+SCH+PAR+SAT) over and above the effect of medication alone (MED). Our nested design will allow us to dismantle the multimodality effect, if one should emerge, in terms of the most common psychosocial (PAR) and an emerging psychoeducational treatment (SCH).